SlideSafe: Development of an Arduino-Based Real-Time Landslide Prediction and

Alert System for Early Warning and Disaster Mitigation

A Research

Presented to

Calbayog City National High School

Calbayog City

In Partial Fulfillment of the Requirements

For Science, Technology and Engineering

Daguman, Loreen Faith Relos, Dze Anne

Delmo, Carl Spencer Talahiban, Shanin Iyhaz
Lacasa, Quinn Rou Talon, Dwight David
Laoyon, Mica Tarrayo, Janiña
Obseñares, Frane Jermaine Tumandao, Hannah Celeste

October 2024

PAGE \* MERGEFORMAT 36
 Introduction

Landslides are one of the most pervasive natural disasters worldwide, resulting

in significant loss of life, property, and infrastructure. With climate change

intensifying weather patterns, the frequency and severity of landslides are expected to

increase globally. Landslides are caused by a combination of factors that contribute to

slope instability. Geological conditions such as steep slopes and loose materials are

significant, as these features can predispose an area to landslides. Hydrological

factors, particularly the saturation of soil due to heavy rainfall or rapid snowmelt, play

a crucial role; studies indicate that when soil saturation reaches around 30% to 50%,

the risk of landslides increases significantly. According to the United Nations Office

for Disaster Risk Reduction (UNDRR), landslides are responsible for thousands of

fatalities each year, particularly in mountainous regions and areas prone to heavy

rainfall (UNDRR, 2020). Between 2004 and 2016, a staggering 55,997 people lost

their lives in 4,862 distinct landslide events globally, with Asia being the most

affected region (Froude, & Petley, 2018). Annual casualties in the United States alone

range from 25 to 50, while worldwide fatalities are typically in the thousands (Xu et

al., 2024). These events create billions of dollars in infrastructural damage,

underscoring the need for effective risk management and prevention strategies

(LANDSLIDES @ NASA., 2023).

Countries around the world, particularly those with mountainous terrains like

Japan, India, and Nepal, have been investing heavily in landslide detection and early

warning systems (JICA, 2018; NIDM, 2019). These systems often involve a

combination of satellite imagery, remote sensing, and ground-based monitoring to

predict landslides and issue timely warnings (Yang et al., 2015). However, in many

PAGE \* MERGEFORMAT 36
developing nations, the deployment of sophisticated technology remains a challenge

due to financial constraints and lack of infrastructure. The use of cost-effective,

accessible technologies like Arduino for landslide detection presents a practical

solution for these regions, helping to save lives and protect communities (Das &

Yadav, 2018).

The Philippines, an archipelago characterized by mountainous landscapes and

heavy rainfall, is highly susceptible to landslides. Typhoons and monsoon rains

frequently trigger landslides, particularly in rural and mountainous regions

(PHIVOLCS, 2021). The government, through agencies like the Philippine Institute

of Volcanology and Seismology (PHIVOLCS) and the Department of Science and

Technology (DOST), has been working to develop and implement early warning

systems to address this hazard (DOST, 2020). However, there remains a significant

need for localised, community-based solutions that can be easily deployed in

vulnerable areas. The integration of Arduino technology into landslide detection

systems offers a low-cost, scalable, and efficient approach to enhancing disaster

preparedness and response at the grassroots level (Cruz et al., 2022).

Calbayog City, located in the province of Samar, is no stranger to the threats

posed by landslides. The region’s topography, coupled with frequent heavy rains,

makes it particularly vulnerable to such disasters (LGU Calbayog, 2019) In recent

years, landslides have caused significant damage to infrastructure, displaced

communities, and even resulted in loss of life (Rappler, 2021). Despite efforts to

improve disaster preparedness, the city still lacks a robust and reliable early warning

system specifically for landslides. This study focuses on the development and

implementation of an Arduino-based landslide detection system tailored to the

PAGE \* MERGEFORMAT 36
specific needs and conditions of Calbayog City. By providing an accessible, real-time

monitoring solution, this project aims to enhance the city's disaster risk reduction

capabilities and contribute to the safety and resilience of its communities (Torres,

2023).

The growing need for effective landslide prediction and early warning

systems has spurred innovations in technology-driven solutions to mitigate their

impact (Smith & Petley, 2009). Existing landslide prediction systems, such as the

GeoSense Landslide Monitoring System, have made strides in early warning and

disaster mitigation. However, these systems often rely on single-sensor data, such as

only using geotechnical sensors or rain gauges, and depend on threshold-based alerts

that may not accurately reflect the complex, multifactorial conditions leading to

landslides. Moreover, such systems tend to be expensive and require significant

technical expertise for deployment and maintenance, making them less suitable for

remote or resource-limited areas.

In contrast, this study proposes the development of an Arduino-based real-

time landslide prediction and alert system. This aims to fill the existing research gap

by integrating multiple sensors, including soil moisture, vibration, and rainfall, to

enhance prediction accuracy. Additionally, it incorporates machine learning

algorithms for predictive analytics and offers a cost-effective, easy-to-deploy solution

designed for sustainability and maintenance in challenging environments.

PAGE \* MERGEFORMAT 36
 Statement of the Problem

This study aims to develop SlideSafe: an Arduino-Based Real-Time Landslide

Prediction and Alert System to provide early warnings of upcoming landslides and

prevent disasters.

Specifically, this study is sought to answer the following questions:

1. How effective is the Arduino-based detection system in predicting landslides in

terms of:

1.1 Accuracy; and

1.2 Alert Timeliness?

2. How efficient is the Arduino-based system in terms of:

2.1 Installation Time and Cost; and

2.2 Maintenance Frequency and Cost?

3. What is the reliability of the power source for the Arduino-based detection

system, and how long can the system operate continuously in terms of:

3.1 Power Source Sustainability; and

3.2 Operational Duration?

PAGE \* MERGEFORMAT 36
 Materials and Methods

This chapter details the research design, time and locale of the study,

instrumentation, validation of instrumentation, respondents of the study, sampling

procedure, data gathering procedure, product development, and data analysis used in

the development of the Arduino-Based Real-Time Landslide Prediction and Alert

System for Early Warning and Disaster Mitigation.

Research Design

In this study, the researchers utilized a developmental research design to

design, develop, and evaluate the Arduino-Based Real-Time Landslide Prediction and

Alert System. Developmental research is defined as the systematic study of designing,

creating, and assessing products to ensure they meet criteria for functionality and

effectiveness (Seels & Richey, 1994). This design was critical for iteratively

developing the system, which involved integrating various sensors, processing units,

and communication modules to create a functional landslide prediction tool.

Additionally, the researchers employed a descriptive research design to evaluate the

perceived efficacy and practicality of the landslide prediction system. Descriptive

research is characterized by observing and describing phenomena as they naturally

occur, without manipulating variables (Prieto et al., 2016). The combination of

developmental and descriptive research designs provided a comprehensive framework

for developing a practical landslide prediction system while simultaneously assessing

its effectiveness and user satisfaction in a real-world context.

Time and Locale of the Study

PAGE \* MERGEFORMAT 36
Research Instrumentation

This study utilized a data collection sheet to obtain the following variables:

Accuracy and Alert Timeliness. A data collection sheet is a systematic tool for

collecting and analyzing data in research. It may also be characterized as a worksheet

that assists the researcher in gathering, processing, and making sense of information

from multiple data resources (Formplus, 2020). The instrumentation is divided into

three (2) parts:

Part I. Accuracy. To evaluate the accuracy, the study focuses on measuring

True Positives, False Positives, and False Negatives, without considering True

Negatives. Key metrics include Precision, which assesses the system's ability to avoid

false alarms, Recall, which evaluates its effectiveness in detecting actual landslides,

and the F1-Score, which provides a balanced measure of both. By analyzing these

metrics, the study aims to determine how well the system predicts real landslides

while minimizing missed detection and false alerts.

A. Error Matrix for Evaluation: Results will be organized using this

simplified structure:

Landslide Occurred No Landslide Occurred

System Alert (Yes) TP: Correct prediction of FP: False alarm (no
landslide landslide)
No Alert (No) FN: Missed prediction No evaluated (No TN)

Table 1: Error Matrix Evaluation for SlideSafe: Development of an Arduino-Based Real-Time Landslide
Prediction and Alert System for Early Warning and Disaster Mitigation

PAGE \* MERGEFORMAT 36
 Table 2: Accuracy Test Sheet for SlideSafe: Development of an Arduino-Based Real-Time Landslide
Prediction and Alert System for Early Warning and Disaster Mitigation

Part II. Alert Timeliness. The table presents an overview of evaluating alert

timeliness. this includes, attempts, data acquisition, processing time, and alert

notification. Multiple attempts will be done to acquire data and the processing time is

being recorded. An alert notification is sent after a successful data acquisition and

processing time is recorded. to further discuss the testing process, multiple attempts

PAGE \* MERGEFORMAT 36
will be made to acquire the data. This could be a simple count. Data Acquisition is the

amount of time successfully collected. It can be a numerical value or time. Processing

time is the duration required to process the acquired data. This would be a time

measurement that is measured by seconds. Alert Notification will be the indicator

whether an alert was successfully sent and will be measured by time.

Table 3: Alert Timeliness Test Sheet for SlideSafe: Development of an Arduino-Based Real-Time Landslide
Prediction and Alert System for Early Warning and Disaster Mitigation

Validation of Instrumentation

Data Gathering Procedure

Product Development

A. Materials for Hardware

PAGE \* MERGEFORMAT 36
 The selection of materials for the hardware development of the Arduino-Based

Real-Time Landslide Prediction and Alert System was guided by the need for

durability, reliability, and suitability for operation in harsh environmental conditions.

High-quality electronic components were chosen to ensure the system's long-term

functionality. The core of the system, the Arduino Mega 2560 R3 microcontroller,

was selected for its versatility and ease of integration with various sensors, including

soil moisture sensors, inclinometer, and weather station model. These sensors were

made from robust materials capable of withstanding exposure to soil, water, and

temperature fluctuations, ensuring accurate and consistent data collection.

To power the system, solar panels were chosen due to their sustainability and

ability to provide continuous energy in remote areas. High-capacity rechargeable

batteries were paired with these panels to store energy and ensure the system's

operation during periods of low sunlight. The materials used for the weatherproof

enclosures were carefully selected to protect the electronic components from

moisture, dust, and temperature extremes. These enclosures were constructed from

durable, non-corrosive materials, ensuring that the sensitive hardware remained

shielded from environmental hazards.

The choice of communication modules, such as GSM specifically SIM900 and

Wi-Fi, was influenced by the need for reliable data transmission over long distances.

These modules were selected based on their proven performance in similar remote

sensing applications. The overall selection of materials was meticulously planned to

ensure that every component of the hardware was built to withstand the rigors of its

PAGE \* MERGEFORMAT 36
intended environment, providing a reliable and effective early warning system for

landslide detection.

B. Materials for Software

The development of the software for the Arduino-Based Real-Time Landslide

Prediction and Alert System required careful selection of programming tools,

libraries, and algorithms to ensure precise data processing, efficient communication,

and reliable alert mechanisms. The core software was developed using the Arduino

IDE, which provides a flexible and user-friendly environment for coding the

microcontroller. This platform was chosen for its compatibility with the Arduino

hardware and its extensive support for sensor integration and communication

modules.

To manage the data collected from various sensors, specific libraries were

utilized for each type of sensor. For example, libraries for the soil moisture sensors,

inclinometer, and digital rain gauge were employed to facilitate accurate data

acquisition and processing. These libraries were selected for their ability to optimize

sensor performance and ensure that the readings were both accurate and timely.

Additionally, the software included signal processing algorithms designed to analyze

sensor data in real-time, allowing the system to detect early signs of potential

landslides based on predefined thresholds.

For communication purposes, the software integrated libraries that supported

GSM and Wi-Fi modules, enabling the system to send alerts via SMS and upload data

to a cloud server for remote

PAGE \* MERGEFORMAT 36
monitoring. The software was also designed to include fail-safes and redundancy

checks to ensure that alerts were transmitted even in the event of intermittent

connectivity.

Finally, the software's user interface was kept simple and intuitive, allowing

for easy configuration and monitoring of the system. This included an LCD display

for local status updates and diagnostic information, as well as LED indicators to

provide immediate visual feedback on system performance. The overall software

architecture was developed with a focus on reliability, efficiency, and ease of use,

ensuring that the system could operate autonomously and deliver accurate and timely

alerts to mitigate the risks associated with landslides.

C. Methods for Hardware

The hardware development of the SlideSafe: Arduino-Based Real-Time

Landslide Prediction and Alert System involved careful selection and integration

of various sensors with the Arduino microcontroller to ensure accurate detection

of landslide indicators. The soil moisture sensor (YL-69) was integrated with the

Arduino's analog input to measure soil saturation levels, providing early warning

when the soil approaches dangerous moisture thresholds. The MPU6050

inlcinometer was connected via the I2C interface to monitor ground vibrations and

shifts in real-time, detecting any unusual movements indicative of terrain

instability. A digital rain gauge was also incorporated to track rainfall, with digital

pulse inputs recording the amount of rain, as excessive rainfall can trigger

landslides. Powering the system, a solar panel and rechargeable battery were used

to ensure continuous operation in remote locations. The hardware components

were enclosed in a weatherproof casing to protect them from environmental

PAGE \* MERGEFORMAT 36
 conditions while ensuring long-term durability and functionality in real-world

scenarios.

D. Methods for Software

The software for the Arduino-based landslide prediction and alert system was

developed using a structured approach to ensure real-time data processing and

accurate prediction. The system’s firmware was programmed in C++ using the

Arduino Integrated Development Environment (IDE). The GPS module provided

location data, which was integrated into the system to correlate sensor readings with

specific geographic coordinates. Data from all sensors were continuously monitored

and processed using predefined threshold values and algorithms to detect anomalies

indicative of potential landslides. When thresholds were exceeded, the system

triggered an alert through the buzzer and displayed real-time data on the I2C LCD.

Additionally, the SD card module was programmed to log all sensor readings and

alert events for further analysis and record-keeping. Debugging and testing were

carried out iteratively, using serial monitoring and data logging to fine-tune the

system’s performance. The software was designed with scalability in mind, allowing

for future integration with wireless communication modules to transmit alerts

remotely.

E. Block Diagram

Arduino
Data Alert
Sensors microcontro
processing system
ller

Communic
ation
PAGE \* MERGEFORMATmodule
36
Testing of the Product

Schematic Diagram

Data analysis

The researchers utilized statistical tools and observations to determine the

efficacy of SlideSafe: Development of an Arduino-Based Real-Time Landslide

Prediction and Alert System for Early Warning and Disaster Mitigation.

Matrix for Accuracy Calculation. This will be utilized to determine the

Accuracy.

Precision: The proportion of correct alerts among all the system’s alerts. It measures

how many of the predicted landslides were accurate.

TP
Precision =
TP+ FP

It helps asses how well the system avoids false alarms.

Recall (Sensitivity): The proportion of the actual landslides that were correctly

predicted by the system.

TP
Recall =
TP+ FN

This metric evaluates the system’s ability to correctly detect all landslides.

F1-Score: The harmonic mean of precision and recall, providing a balanced view of

the system’s performance when both false alarms and missed events are important.

PAGE \* MERGEFORMAT 36
 Precision x Recall
F1-Score = 2 x
Precision+ Recall

Mean. This will be utilized to calculate the alert timeliness. It will provide the

average time taken for the data acquisition process across all attempts offering

insights into the systems efficiency and helping identify any delays or areas for

optimization.

PAGE \* MERGEFORMAT 36
 List of References

UNDRR (2020):
United Nations Office for Disaster Risk Reduction. (2020). Global Assessment
Report on Disaster Risk Reduction 2020.
Retrieved from [[Link]

Smith, M., & Petley, D. N. (2009):

Smith, M., & Petley, D. N. (2009). Landslide Hazards, Risks, and Disasters.
Academic Press.

JICA (2018):
Japan International Cooperation Agency. (2018). Landslide Risk Reduction in Asia-
Pacific Regions.
Retrieved from [[Link]

NIDM (2019):
National Institute of Disaster Management (NIDM). (2019). Landslide Risk
Mitigation Strategies in India.
Retrieved from [[Link]

Yang et al. (2015):

Yang, X., Hong, Y., & Chen, Y. (2015). *Landslide Susceptibility Mapping and
Early Warning Using Remote Sensing and GIS Technologies*. Journal of
Environmental Management, 150, 189–200.

Das, S., & Yadav, M. (2018):

Das, S., & Yadav, M. (2018). A Cost-Effective Arduino-Based Landslide Detection
System. IEEE Access, 6, 3156-3163.
DOI:
[10.1109/ACCESS.2018.2791550]([Link]

PHIVOLCS (2021):
Philippine Institute of Volcanology and Seismology (PHIVOLCS). (2021).
Landslide Hazard Information and Mapping in the Philippines.
Retrieved from
[[Link]

DOST (2020):
Department of Science and Technology (DOST). (2020). Disaster Risk Reduction
Initiatives in the Philippines.
Retrieved from [[Link]

PAGE \* MERGEFORMAT 36
Cruz, M., Bautista, M., & Santos, R. (2022):
Cruz, M., Bautista, M., & Santos, R. (2022). Using Arduino Technology for
Disaster Risk Reduction in Rural Communities.
Journal of Environmental Science and Technology, 14(3), 198-210.
DOI: [10.1016/[Link].2021.12.004]([Link]

LGU Calbayog (2019):

Local Government Unit of Calbayog. (2019). Landslide Hazard and Disaster
Preparedness in Calbayog City.
Retrieved from [[Link]

Rappler (2021):
Rappler. (2021). Landslide Kills 3, Displaces Dozens in Calbayog City.
Retrieved from [[Link]

Torres, A. (2023):
Torres, A. (2023). Building Resilience: A Community-Based Landslide Detection
System for Calbayog City.
Philippine Journal of Natural Disaster Science, 45(2), 124-135.

PAGE \* MERGEFORMAT 36